On Thin Ice: Does Uniform Color
Really Affect Aggression in
Professional Hockey?
Social Psychological and
Personality Science
000(00) 1-5
ª The Author(s) 2010
Reprints and permission:
sagepub.com/journalsPermissions.nav
DOI: 10.1177/1948550610389824
http://spps.sagepub.com
David F. Caldwell1 and Jerry M. Burger1
Abstract
Past research suggests that wearing either a black or a red uniform leads to increased aggression or an increase in perceived
aggression during professional sports. However, this research suffers from a number of limitations, including an inability to
manipulate the independent variable. A recent change in the National Hockey League’s uniform policy created the possibility
of a naturally occurring experiment that allowed the authors to examine whether aggression levels were higher when teams
wore black or red jerseys. The authors compared games against the same opponent in which home teams wore red or black
jerseys for one game and their usual color for another game on several measures of aggression. They found no evidence that
either black or red uniforms were related to higher levels of aggression in professional hockey games.
Keywords
aggression, violence, evolutionary psychology, motivation and performance, research methods
Of the many variables researchers have identified as contributing
to aggression, perhaps the most intriguing is the color worn by the
aggressor. To test the effect of color in a real-world setting,
researchers have examined the number of aggressive acts
performed by athletes during sporting events. Findings from these
studies suggest that athletes who wear either black or red may be
more aggressive than athletes who wear other colors.1
Investigators examining the relation between black uniforms
and aggression point to the widely recognized association
between black and badness (Frank & Gilovich, 1988). We commonly speak of black moods and the bad guys wearing black
hats. Research on the affective meanings of colors finds that
black is associated with death and evil in virtually all cultures
(Adams & Osgood, 1973). Because in most contexts aggression
is considered a negative behavior, black also is associated with
aggression. Research tying red uniforms to aggression is often
based in biological and evolutionary concepts. In particular,
biologists have identified the importance of red in intraspecies
competition. In many species, the amount or intensity of red is
associated with status and dominance (Pryke, Andersson,
Lawes, & Piper, 2002; Setchell & Wickings, 2005). During
competitive interactions between males, red is typically indicative of dominance, which gives the male who expresses more or
more intense red an advantage. In humans, we commonly associate a red face with anger, a face drained of color with fear.
and hockey players. Over a 16-year period, they found
consistent evidence that teams wearing black uniforms were
penalized more often than teams wearing nonblack uniforms.
The researchers also point to two hockey teams that switched
from nonblack to black uniforms during this period. In both
cases, the teams were penalized more often after the switch
than before. The investigators correctly point out the referees’
perception also might have been affected by the players’
uniform color. Thus, whether the findings reflect the actual
number of fouls or the referees’ perception that the team in
black was committing more fouls is difficult to tease apart.
In follow-up studies, Frank and Gilovich found evidence that
both of these processes might be operating.
Hill and Barton (2005b) examined the effect of red uniforms
in four combat sports (boxing, tae kwon do, Greco-Roman
wrestling, freestyle wrestling) during the 2004 Olympics. The
two contestants in each round of these competitions were randomly assigned to wear either red or blue protective gear. The
researchers found that the athlete wearing the red uniform was
victorious significantly more often than could be attributed to
chance. In another investigation, these researchers also found
that English football teams wearing red jerseys had better
1
Uniform Color and Aggression
Frank and Gilovich (1988, Study 2) examined the relation
between uniform color and penalties in professional football
Santa Clara University, Santa Clara, CA, USA
Corresponding Author:
Jerry M. Burger, Santa Clara University, Department of Psychology, 500 El
Camino Real, Santa Clara, CA 95053
Email: [email protected]
Downloaded from spp.sagepub.com at Santa Clara University on February 8, 2016
2
Social Psychological and Personality Science 000(00)
won–loss records than teams wearing other colors (Attrill,
Gresty, Hill, & Barton, 2008).
Several hypotheses have been put forth to account for the
advantage of wearing red uniforms. It may be that the athlete
responds to the color he or she is wearing and performs more
aggressively (Hill & Barton, 2005a, 2005b). But it might also
be the case that seeing an opponent in red disrupts the performance of the athlete not wearing red (Elliot, Maier, Moller,
Friedman, & Meinhardt, 2007). Alternatively, red may simply
be easier or more difficult for opponents to see, which can
affect performance (Rowe, Harris, & Roberts, 2005). Finally,
as with football, it may be the case that judges are affected
by the uniform color and may award points differently for competitors in red and nonred uniforms (Hagemann, Strauss, &
Leissing, 2008).
Limitations of the Extant Research
The notion that professional athletes either act more aggressively or are seen as acting more aggressively when they wear
certain colors is intriguing and has theoretical implications for
the relation among external stimuli, identity, and behavior
(Frank & Gilovich, 1988). At a practical level, if the effect is
powerful enough to influence the outcomes of important sports
events like the Olympics, steps should be taken to neutralize
the unfair advantage given to some competitors. However, a
close examination of the studies demonstrating the effect suggests that the evidence for a causal link between uniform color
and aggression might not be as strong as many believe.
In fact, there are several limitations in this research. First,
the study that found more aggression in football and hockey
players when the players wore black uniforms did not manipulate the independent variable (Frank & Gilovich, 1988). That is,
whether a team wore black or some other color was not random. Professional sports teams are free to select their own uniform colors. No doubt many factors come into play when
making this decision, but key among these reasons is the kind
of image owners and management want to portray. Indeed, professional sports has become a multi-billion-dollar enterprise,
and many of the issues surrounding branding and marketing
that concern other large businesses most likely are important
to the sports industry as well. Selecting black for the team color
is no accident. If, as Frank and Gilovich (1988) argue, black has
a strong association with intimidation and aggression, then
making black the team color is a strong indication of the general style of play owners and management want and expect out
of their athletes. This expectation is likely to affect team performance in many ways. Most obviously, teams trying to develop
a reputation for toughness and aggression can select players
with matching reputations. The message also can be delivered
through advertising and other team promotions that feature
aggressive attitudes and play. Perhaps most important, players
can be encouraged directly or indirectly by owners, coaches,
teammates, and perhaps even fans to play more aggressively.
Through these and other channels, a team that wants to promote
an aggressive image is likely to practice and play within an
aggressive team culture. As a result, without random
assignment, it is impossible to know whether wearing black
uniforms makes teams more aggressive (or makes them appear
more aggressive to referees) or whether teams that want to be
more aggressive select black uniforms.
Second, the studies linking red uniforms and aggression did
not measure aggression directly. Rather, the investigators
examined how often the red-clad competitor won in a combative sport (e.g., wrestling). Although increased aggressiveness
may heighten the chances of winning, many other skills, tactics, and attributes also affect the outcome of these matches.
Among other possibilities, competing against someone in a red
uniform might be disruptive (Elliot et al., 2007) or intimidating
(Ioan et al., 2007). The extent to which the red uniform is visible against the background could also account for the effect
(Rowe et al., 2005). Finally, more aggression does not necessarily mean more success. It could be the case that being too
aggressive sometimes hurts performance in these sports.
Third, research examining uniform colors may suffer from
confounds because of the location of the event and the competition. In team sports, uniform color is almost always confounded with whether a game is played at home or away.
Typically in professional football and hockey, the home team
wears the team color and the visiting team wears white. This
is particularly problematic in that home teams win a higher percentage of games. In American professional sports, the percentage of games won by home teams during the 2009 season
ranged from 54.8% in baseball to 59.8% in basketball. There
also is evidence that the home team may be called for fewer
fouls than the visiting team (Anderson & Pierce, 2009). In addition, in some sports like football or judo, competitors may play
each other only once, which makes it impossible to eliminate
effects from specific matchups between opponents or the
dynamics of a particular game (Timmerman, 2007).
Fourth, the dependent measure used in the research examining aggression in football games is a concern. Frank and
Gilovich (1988) used the total number of yards the team was
penalized as their measure of aggression. They correctly point
out that aggressive penalties tend to draw more yards than less
aggressive penalties. However, overly aggressive play is but
one reason for being penalized in football. The most common
penalties (e.g., false starts, holding, offsides), and even some
of the most costly penalties (e.g., pass interference) are not very
good indicators of aggression. Thus, a better measure of
aggression for football players would look only at penalties that
clearly reflect aggressive acts. This limitation is less of a concern for the hockey data because hockey penalties are primarily
the result of overly aggressive play.
The Present Study
Conclusively determining whether uniform color affects
aggressiveness or perceived aggressiveness requires a study
with either random assignment of uniform or a withinsubjects design in which a team wears different colors on different occasions. Moreover, the study needs to look at clear and
Downloaded from spp.sagepub.com at Santa Clara University on February 8, 2016
Caldwell and Burger
3
direct indicators of aggressive behavior and eliminate
confounding factors such as where the game was played and who
the opponent was. Fortunately, a recent change in the National
Hockey League (NHL) uniform policy provides an opportunity
to conduct a study that meets all of these requirements.
As in most sports, professional hockey has an established
protocol for uniform colors. Traditionally each team has two
uniforms. One uniform has a dark jersey in the team’s color and
is normally worn for home games. The other uniform has a
white jersey and is normally used for away games. However,
during the 2008–2009 and 2009–2010 seasons, the NHL
allowed teams to wear a third jersey for a limited number of
home games. Some teams used their existing home color and
simply changed the logo to create the third jersey. However,
other teams selected a different color for their third jersey.
Teams using a third jersey wore that jersey for a limited number of prespecified home games.
Among the teams that used a different color for their third
jersey, eight selected the color black. This development created
a situation in which eight teams wore both their traditional
colored jerseys and, on limited occasions, black jerseys for
home games. For example, the San Jose Sharks regularly wore
black jerseys for games played on Tuesdays or Thursdays and
their traditional teal colored jerseys for games played on other
days. Moreover, because NHL teams may play the same opponent at home more than once in a season, some teams played an
opponent one time wearing black jerseys and another time
wearing their normal home colored jersey during the same season. A similar situation developed for red jerseys. Six teams
selected a third jersey that created a situation in which sometimes they wore a red jersey (either as their primary home jersey or as the third jersey) and sometimes they wore a different
colored jersey for home games. Again, over the course of a season, these teams could play the same opponent at home once
wearing a red jersey and once wearing a nonred jersey.
The use of the third jersey allows for a repeated measures
design with the unit of analysis being a pair of games. That is,
we can compare the same two teams playing in the same arena
in which the home team wore black or red for one game and a
nonblack or nonred color for the other game. The visiting team
wore light colored jerseys for both games. This design eliminates many of the problems found in earlier studies. The jersey
color is not confounded by the location of the game or the specific opponent. Since teams wear their third jerseys throughout
the season, there is no systematic order effect for jersey color.
Using pairs of games in the same season also ensures that the
vast majority of players will be the same in both games and that
a change in team colors does not reflect a new team strategy.
Finally, professional ice hockey is a particularly good vehicle for testing the effects of color on aggression. Most of the
penalties in hockey are for aggressive behavior (e.g., roughing,
spearing, slashing), and fighting is an accepted part of the
game. Conn Smythe, a long-time player and coach for whom
the playoff most valuable player award is named, summed up
his philosophy by remarking, ‘‘If you can’t beat’em in the
alley, you can’t beat’em on the ice’’ (Levin, 2006).
Method
We reviewed NHL team web sites to identify teams that wore
both a black (or red) jersey and a different colored jersey for
home games during the 2008–2009 and 2009–2010 seasons.
We obtained a list of the dates when the third jersey was worn
and reviewed season schedules to determine whether the home
team also played the same opponent that season while not
wearing the third jersey. Over the two seasons, there were
102 pairs of games in which the home team played the same
opponent within a single season while once wearing black and
once wearing nonblack jerseys.2 The two games in the pair
were separated by an average of 24.65 games (SD ¼ 17.01),
and in 46% of the pairs the black jersey was worn in the first
game. After identifying the relevant games, we used official
score sheets to obtain the final score and number and type of
penalties in each game. We followed the same process to create
pairs of games in which the home team wore red and another
color of jersey. We identified 61 pairs of games over the two
seasons. The two games were separated by an average of
24.16 games (SD ¼ 16.71), with red jerseys being worn in the
first game 57% of the time.
We examined three fairly unambiguous measures of aggression and two indirect measures of aggression. First, we looked
at the total number of penalty minutes assessed to the team
(most infractions carry a 2-minute penalty). We conducted this
analysis for both the home team and the visiting team. Second,
we looked at the number of severe penalties. We created a composite measure by summing the number of penalties for three
particularly violent infractions, roughing (a punch or punching
motion normally directed to the head), fighting (continued
punching), and game misconduct (normally given if a penalized behavior leads to an injury). Again, we examined this
measure for both the home and visiting teams. Third, we looked
at the number of games that became atypically violent, as indicated by the total number of penalty minutes assigned to both
teams. Although not necessarily an indicator of aggression, like
some earlier investigators, we also examined two performance
measures. First, we looked at whether the home team won the
game. Second, in the NHL, teams earn two points for a win, one
point for a loss in overtime, and no points for a loss in regulation play. We looked at how often the home team earned at
least one point.
Results
The Effect of Black Jerseys
We began by looking at the total number of penalty minutes per
game. In games in which black jerseys were worn, the home
team was penalized an average of 13.70 minutes (SD ¼
10.0). This did not differ significantly from the games in which
the home team wore its regular jerseys (M ¼ 13.12, SD ¼ 7.63),
t(101) ¼ 0.47, p ¼ .64. A similar pattern was found for visiting
teams. Visiting teams were penalized an average of 14.74 minutes (SD ¼ 8.94) when the opponent wore black and 13.83 minutes (SD ¼ 7.42) when it did not, t(101) ¼ 0.84, p ¼ .40.
Downloaded from spp.sagepub.com at Santa Clara University on February 8, 2016
4
Social Psychological and Personality Science 000(00)
Because hockey is generally an aggressive sport, it may be
the case that uniform color affects only the frequency of
extreme acts of aggression. To test this possibility, we also
examined the number of times teams were penalized for severe
instances of aggression. We counted the total number of roughing, fighting, and misconduct penalties for each team per game.
Because these penalties are relatively rare, we used a nonparametric Wilcoxon signed ranks test for related measures for
these analyses. Overall, the results were similar to those for
penalty minutes. When wearing black, the home team received
an average of 1.38 severe penalties (SD ¼ 1.76) compared to
1.33 severe penalties (SD ¼ 1.40) when wearing its standard
jersey. This difference is not significant (p ¼ .97). There also
was no significance difference in these penalties for the visiting
team when the opponent wore black (M ¼ 1.36, SD ¼ 1.63)
versus when the home team wore its standard jerseys (M ¼
1.24, SD ¼ 1.32; p ¼ .45). Controlling for the time between
games and the order of the games did not change the results.
Even though jersey color is unrelated to penalties and severe
penalties, it is possible that when one team wears black, the
general level of aggression is prone to escalation. That is,
aggression may get out of hand during some games, resulting
in an excessive number of penalties. To test this, we calculated
the total penalty minutes for all 204 games (M ¼ 27.69, SD ¼
15.73) and identified games in which total penalty minutes
were one standard deviation or more above the mean. Of the
games 15% exceeded this threshold when the home team wore
black, compared to 11% of the games in which black was not
worn. The difference between these proportions is not significant (z ¼ 1.10).
Finally, we also examined whether home teams were more
successful when they wore black jerseys rather than their normal jerseys. When home teams wore black, they won 57% of
their games and scored points (either a victory or an overtime
loss) in 67% of the games. When home teams wore their standard jersey, they won 50% of the time and scored points in 60%
of the games. The color of jersey worn by the home team was
not significantly related to the proportion of wins, w2(1, N ¼
102) ¼ 1.11, p ¼ .58, or games in which points were received,
w2(1, N ¼ 102) ¼ 1.06, p ¼ .60.
The Effect of Red Jerseys
We used the same analytic strategy to compare the effect of red
jerseys. In games in which the home team wore red, the home
team was penalized an average of 12.02 minutes (SD ¼ 6.52).
When the home team wore a nonred jersey, it was penalized
an average of 13.30 minutes (SD ¼ 9.82). The difference is
not statistically significant, t(60) ¼ 0.97, p ¼ .34. Similarly,
the visiting team was penalized an average of 13.39 minutes
(SD ¼ 8.61) when the home team wore red jerseys and 15.07
minutes (SD ¼ 8.75) when the home team wore another
color. This difference also is not statistically significant,
t(60) ¼ 1.25, p ¼ .21.
When the home team wore red, it committed an average of
1.25 severe penalties (SD ¼ 1.37). When the team wore nonred
jerseys, the average number of these penalties was 1.30 (SD ¼
1.77). When the visiting team played against a team wearing
red jerseys, it committed an average of 1.23 severe penalties
(SD ¼ 1.24). The visiting team committed an average of 1.44
of these penalties (SD ¼ 1.73) when the opponent did not wear
red. Wilcoxon signed ranks tests for related measures showed
no significant differences in the severe penalties committed
in the two games by the home team (p ¼ .72) or the visiting
team (p ¼ .75).
We again looked at whether the number of highly penalized
games differed depending on the home team jersey color. The
average number of total penalty minutes across all 122 games
was 26.89 (SD ¼ 15.21). When the home team wore red, 8% of
the games had a penalty total that was at least one standard
deviation above the mean. When the home team did not wear
red, 15% of the games met this threshold. This difference is not
significant (z ¼ 1.50).
Finally, when the home team wore red, it won 47% of the
games and scored points in 62%. When the home team wore
a nonred jersey, it won 56% of the time and earned points in
61% of the games. The color of jersey worn by the home team
was not significantly related to the proportion of wins, w2(1,
N ¼ 61) ¼ 0.99, p ¼ .62, or the proportion of games in which
points were received, w2(1, N ¼ 61) ¼ 0.11, p ¼ .74.
Discussion
The notion that uniform color causes aggression in sports is an
intriguing one. It has been used to support the ideas that cultural
artifacts (e.g., ‘‘black hat’’) can shape behavior or that mechanisms observed in other species (e.g., status systems among red
widowbirds) can influence complex human actions. Although
these ideas are provocative, making unequivocal statements
about the connection between uniform color and aggression
seems at best premature.
The pattern in the data we collected is remarkably consistent. No matter how we measured aggression—total penalty
minutes, number of severe penalties, likelihood that a game
would have an excessive number of penalty minutes, and team
success—we found no evidence that wearing either a black or a
red jersey increases the amount of aggression in hockey
games.3 This research represents more than a failure to replicate. Rather, it is the first study that eliminates or controls for
several limitations that plagued earlier investigations. We
examined clear measures of aggression, controlled for game
location and opponent, and compared games in which the uniform color was randomly assigned. At least two other investigations also have failed to find support for the notion that
athletes wearing black uniforms are more aggressive (Mills
& French, 1996; Tiryaki & Scedilfik, 2005). However, these
studies also had some of the methodological limitations as the
studies reporting an effect for color.
Of course, failure to find a connection between uniform
color and aggression does not mean the effect does not exist.
It is possible that researchers who control for the limitations
we identify might yet find the effect when looking at different
Downloaded from spp.sagepub.com at Santa Clara University on February 8, 2016
Caldwell and Burger
5
sports or perhaps when using different measures of aggression.
Nonetheless, until more data are available, the most appropriate
conclusion from the literature seems to be that there is little evidence either that wearing black or red uniforms causes athletes
to act more aggressively or that athletes wearing these colors
are perceived by officials as being more aggressive.
Declaration of Conflicting Interests
The authors declared no potential conflicts of interests with respect
to the authorship and/or publication of this article.
Financial Disclosure/Funding
The authors received no financial support for the research and/or
authorship of this article.
Notes
1. Rowe, Harris, and Roberts (2005) report that competitors randomly
assigned to wear blue rather than white during the 2004 Olympics
judo competition won more often than would be expected by
chance. However, Dijkstra and Preenen (2008) found that after
controlling for confounding factors, such as the time between
matches and seeding in those matches, the effect disappears.
2. If a home team played the opponent once while wearing the third
jersey and more than once wearing the normal jersey, we selected
the game closest in time to create the pair.
3. The failure to find significant effects cannot be easily attributed to
a lack of power. The sample size for games with black jerseys and
for games with red jerseys is within the recommended range for
uncovering medium-sized effects (Cohen, 1992).
References
Adams, F., & Osgood, C. (1973). A cross-cultural study of the
affective meanings of color. Journal of Cross-Cultural Psychology, 4, 135-156.
Anderson, K., & Pierce, D. (2009). Officiating bias: The effect of foul
differential on the subsequent fouls in NCAA basketball. Journal
of Sport Sciences, 27, 687-694.
Attrill, M., Gresty, K., Hill, R., & Barton, R. (2008). Red shirt colour
is associated with long-term success in English football. Journal of
Sports Sciences, 26, 577-582.
Cohen, J. (1992). A power primer. Psychological Bulletin, 112,
155-159.
Dijkstra, P., & Preenen, P. (2008). No effect of blue on winning contests in judo. Proceedings of the Royal Society: Biological
Sciences, 275, 1157-1162.
Elliot, A. J., Maier, M. A., Moller, A. C., Friedman, R., & Meinhardt, J.
(2007). Color and psychological functioning: The effect of red on
performance attainment. Journal of Experimental Psychology:
General, 136, 154-168.
Frank, M., & Gilovich, T. (1988). The dark side of self- and social perception: Black uniforms and aggression in professional sports.
Journal of Personality and Social Psychology, 54, 74-85.
Hagemann, N., Strauss, B., & Leissing, J. (2008). When the referee
sees red. Psychological Science, 19, 769-771.
Hill, R., & Barton, R. (2005a). Hill and Barton reply. Nature, 435,
E10-E11.
Hill, R., & Barton, R. (2005b). Red enhances human performance in
contests. Nature, 435, 293.
Ioan, S., Sandulache, M., Avramescu, S., Ilie, A., Neascu, A., Zagrean,
L., & Moldovan, M. (2007). Red is a distractor for men in competition. Evolution and Human Behavior, 28, 285-293.
Levin, A. (2006). Cure that puckish feeling at Toronto’s shrine to
hockey. Psychiatric News, 41, February 17, 21.
Mills, B., & French, L. (1996). Assertive behavior and jersey color in
the national hockey and football leagues: The black uniform myth
and reality. Journal of Human Movement Studies, 31, 47-60.
Pryke, S., Andersson, S., Lawes, M., & Piper, S. (2002). Carotenoid
status signaling in captive and wild red-collared widowbirds: Independent effects of badge size and color. Behavioural Ecology, 13,
622-631.
Rowe, C., Harris, J., & Roberts, C. (2005). Seeing red? Putting sportswear in context. Nature, 435, E10.
Setchell, J., & Wickings, E. (2005). Dominance, status signals, and
coloration in male mandrills (Mandrillus sphinx). Ethology, 111,
25-50.
Timmerman, T. (2007). ‘‘It was a thought pitch’’: Personal, situational, and target influences on hit-by-pitch events across time.
Journal of Applied Psychology, 92, 876-884.
Tiryaki, M., & Scedilfik, S. (2005). Assessing whether black uniforms
affect the decisions of Turkish soccer referees: Is the finding of
Frank and Gilovich’s study valid for Turkish culture? Perceptual
and Motor Skills, 100, 51-57.
Bios
David F. Caldwell is the Stephen and Patricia Schott Professor
of Management in the Leavey School of Business at Santa
Clara University.
Jerry M. Burger is a Professor of Psychology at Santa Clara
University.
Downloaded from spp.sagepub.com at Santa Clara University on February 8, 2016